Aging is associated with an increased risk for a number of diseases, including colorectal cancer (CRC). An accumulation of undesirable cellular changes over time is known to increase CRC risk. Changes in DNA methylation, an epigenetic mechanism that exerts control over gene expression, occur with age and have also been associated with many types of cancer. The age-related DNA methylation changes are commonly referred to as ‘methylomic drift’ but whether this drift is directly involved in the development of CRC is not fully understood. Furthermore, a precise understanding of the timing between the occurrence of cellular insults that initiate a founder premalignant cell to later development of clinically detectable CRC has not been well defined.
Drs. Georg Luebeck and William Hazelton and colleagues in the Divisions of Public Health Sciences and Clinical Research conducted a study to address these questions and reported their results in a recent issue of the journal Cancer Research. Previous work by the authors on Barrett’s esophagus (BE) and tissue aging set the stage for the new study (Curtius et al., 2016; Luebeck et al., 2017). Dr. Luebeck explained, “We demonstrated that the presence of differential methylomic drift in BE tissue vs normal esophageal tissue allowed us to potentially determine how long a patient has lived with BE, which does not cause symptoms when and after it arises in the esophagus. Because age related changes in cells (e.g. increased mutational burden, impairment of the immune system, inflammation, etc) are likely involved in the pathogenesis of many cancers, we believe an accurate prediction of the duration BE is present in a person may improve our cancer risk predictions for patients with BE.”
In the new study, the authors assessed DNA methylation at the genome-wide level in normal and neoplastic colorectal tissues. In normal tissues, age-associated methylomic drift occurred across the genome, affecting thousands of CpG loci (DNA sequences where methylation can occur) in hundreds of CpG islands (regions with high density of CpG loci) genome-wide. To determine whether drift is altered in CRC, the authors compared drift rates in normal and neoplastic tissues. CpG islands that demonstrated methylomic drift in normal tissues also exhibited drift in the cancer tissues but at a significantly higher rate, estimated to be three to four times that in normal tissues. However, CpG islands that do not drift with age in normal tissue were found to also not drift discernibly in cancer tissue, although they may be altered.
Image provided by Dr. Georg Luebeck
The authors then incorporated the methylomic drift data into modeling analyses to calculate expected sojourn times from the birth of a founding premalignant cell to the time of detectable CRC. Overall, the modeling estimated that a founder cell is likely born before the age of 20. “These results suggest that the founding premalignant cell that eventually becomes a cancer is frequently initiated early in life, undergoes slow premalignant growth as a micro-adenoma and adenoma, and may sojourn for decades before malignant conversion, cancer growth, and detection by symptoms or through cancer screening (See Figure),” said Dr. Hazelton. This striking finding suggests that the modification of environmental and lifestyle factors that influence CRC risk may have the greatest impact if introduced early in life.
With an important role of DNA methylation in the regulation of gene expression, the authors also probed whether methylomic drift was associated with repression of gene expression in CRC tissues using data collected from The Cancer Genome Atlas (TCGA). In the right colon, 56% of the identified gene/CpG island pairs were significantly negatively correlated, indicating methylation repressed gene expression, while only 5% of the pairs were positively correlated, indicating methylation upregulated gene expression. Many fewer gene/CpG island pairs were significantly associated in the left colon, but the overall trend was similar, with a greater proportion of significant pairs negatively associated (26%) compared to positively associated (7%).
Looking forward, Dr. Luebeck commented, “One of the provocative questions that emerged is whether methylomic drift is simply a 'molecular clock' phenomenon, or something that ultimately exerts selective pressure on normal tissue -and possibly neoplastic tissues with functional consequences that drive the pathogenesis of colon cancer. In other words, is it possible that advanced drift leads to transcriptional deregulation, phenotypic diversity, (epi)genomic instability, and possibly defective cellular reprogramming that determines the malignant state.” The new study results provide early insight to this question, “This possibility is supported by our finding that much of the methylomic drift seen in normal colorectal tissues is associated with significantly altered gene expression in colorectal tumors from TCGA, interestingly much more so in right colon vs left colon,” added Dr. Luebeck.
Fred Hutch/UW Cancer Consortium members Georg Luebeck, Wynn Burke, Paul Lampe, Christopher Li, Polly Newcomb, Andrew Kaz, John Inadomi, and William Grady contributed to this research.
This research was supported by the National Institutes of Health.
Luebeck GE, Hazelton WD, Curtius K, Maden SK, Yu M, Carter KT, Burke W, Lampe PD, Li CI, Ulrich CM, Newcomb PA, Westerhoff M, Kaz AM, Luo Y, Inadomi JM, Grady WM. 2018. Implications of epigenetic drift in colorectal neoplasia. Cancer Research. doi: 10.1158/0008-5472.CAN-18-1682
Curtius K, Wong C-J, Hazelton WD, Kaz AM, Chak A, Willis JE, et al. 2016. A molecular clock infers heterogeneous tissue age among patients with Barrett’s esophagus. PLoS Computational Biology. 12(5), e1004919.
Luebeck EG, Curtius K, Hazelton WD, Maden S, Yu M, Thota PN, et al. 2017. Identification of a key role of widespread epigenetic drift in Barrett’s esophagus and esophageal adenocarcinoma. Clinical Epigenetics. doi: 10.1186/s13148-017-0409-4